It’s worth taking a moment to focus on just one number. Over $20 billion was invested in educational technology by venture capital in 2021. The idea that software could do what schools had failed to do—reach every student, in every country, regardless of geography, income, or the caliber of the person in front of the room—was pursued with twenty billion dollars in a single year. The pitch was spotless. As it happened, the math was not.
EdTech investment fell to about $3 billion by 2024. It’s not a correction. It’s a reckoning. In one of the most spectacular corporate collapses in recent memory, Byju’s, the Indian tutoring behemoth that was once valued at $22 billion—a figure that genuinely thrilled Palo Alto investors—imploded. Once a mainstay of American college life, Chegg saw its user base shift almost instantly to ChatGPT, which offered the same service for free. It turned out that the boom during the pandemic had been a sugar spike. The ensuing crash was brutal.
Nevertheless, the global teacher shortage was worsening during the same period that EdTech was consuming all that money, creating all those dashboards, and creating all those gamified lessons. The percentage of school principals who reported teacher shortages increased from 29% in 2015 to almost 47% in 2022, according to the OECD. By 2030, there will be 44 million fewer primary and secondary teachers worldwide, according to UNESCO. That was not resolved by any adaptive learning algorithm. Not one.
Key Information: The EdTech Industry
| Sector | Education Technology (EdTech) |
| Peak Investment Year | 2021 — over $20 billion in venture capital |
| Investment by 2024 | ~$3 billion — a decade low |
| Decline from Peak | 89% drop in EdTech VC funding (2021–2024) |
| Notable Collapse | Byju’s — once valued at $22 billion, now effectively defunct |
| Other Casualty | Chegg — cratered as students migrated to ChatGPT |
| Global Teacher Shortage | UNESCO projects 44 million teacher vacancies by 2030 |
| US School EdTech Spend (2024) | $30 billion — roughly 10x spending on textbooks |
| Education Sector Projection | Expected to reach nearly $10 trillion globally by 2030 |
| Key Critics | UNESCO, OECD, educators, and now — investors themselves |
The unsettling reality at the heart of this is that EdTech was addressing an issue that education didn’t genuinely have. Delivery of content was never a problem. Instructors were. While the water was running low, the industry spent ten years developing better pipes. From the outside, it seems that the people writing the checks saw schools more as markets than as institutions, places to be disrupted than understood.
The instruments themselves weren’t always flawed. AI can spot learning gaps, identify students who are having difficulty, cut down on administrative burden, and free up teachers’ time to do what only they can do. That is true. However, the tool became the focal point somewhere between the pitch deck and the whiteboard. A good example would be gamified language apps, which are genuinely effective at maintaining user engagement, which is different from teaching someone a language. According to a recent survey, 30% of language app users specifically stopped using them because they weren’t improving. They had become quite proficient with the app. They were still unable to communicate.
EdTech kept creating the wrong things for a structural reason. Engagement can be quantified. Learning is slow and difficult to measure—real learning that endures, transfers, and manifests itself years later in a career or a decision. Metrics, monthly actives, retention curves, and the type of growth that appears in a slide deck are what drive venture capital. Slides with deep pedagogical outcomes don’t look good. As a result, the industry quietly ignored what it couldn’t measure and optimized for what it could. To put it simply, too many platforms prioritize engagement over real results, creating tools that appear impressive but fall short of producing significant educational outcomes, according to Wall Street English CEO James McGowan. He is correct, and it’s noteworthy that the observation is made by someone who is actively using the area rather than observing it from a distance.
The students who were most in need of EdTech’s promise were also the ones who benefited the least from it. Children in underfunded schools in rural Texas or rural India, which were frequently highlighted in pitch decks, frequently had the least dependable internet, the fewest devices, and the most overworked teachers. There was no closure of the digital divide. It widened in a few places. In 2024, US schools spent $30 billion on educational technology—roughly ten times more than they did on textbooks. The actual results of that money are still unknown. The data from the OECD is not promising.
The next stage of EdTech might be more astute in this regard. With $51 billion invested in the field in 2024 alone, AI investment is still growing, and some of it is making its way into educational platforms that, at the very least, are attempting to collaborate with teachers rather than work around them. The distinction is very important. A gamified app designed to maximize screen time is completely different from an AI tool developed with educators, stress-tested against actual classroom conditions, and accountable to quantifiable learning outcomes. In contrast to 2021, the industry is now aware of this. The question remains whether it takes action based on that knowledge.
The teacher in the classroom is not a content bottleneck that can be circumvented, as the past ten years have shown with the kind of clarity that only a very costly failure can produce. The intervention is her. Nothing on a server has been able to replicate the relationship, the judgment, or the ability to read a class of thirty students and determine which three are having difficulty and why. It is not because technology is ineffective in education that the $20 billion EdTech bubble failed. It failed because its designers neglected to inquire about the true nature of education.
